ABSTRACT
The advent of unmanned aerial systems (UASs) permits data collection at high spatial and temporal resolutions, enabling high-throughput phenotyping and precision agriculture. Numerous scientific and technical challenges present themselves associated with methods of UAS data collection, preprocessing, information extraction and synthesis, and decision support. Multiple UAS experiments recently conducted at the Texas A&M University Research Farm near College Station, Texas, highlight issues such as sensor-platform modularity, data acquisition strategies, and preprocessing and information extraction problems that need to be more efficiently addressed. Radiometric and geometric calibration are shown to be both nontrivial and essential for reliable vegetation assessment and monitoring. Various vegetation indices and thematic mapping approaches exist but are currently incapable of diagnostic assessment of plant species or stress status. Evaluation of multiple crop and weed species with spectral and spatial wavelet analysis reveals patterns that may be diagnostic for assessing plant and canopy structure. Ultimately, UASs have great potential for obtaining imagery that can assist in improving crop yield and providing information for decision support, but further research is needed to optimize UAS engineering configurations and to address a variety of issues related to preprocessing of imagery, diagnostic analysis of environmental and plant conditions, and artificial intelligence for decision support.
